Original Article

Validity Of Rockall Scoring System In Determining In- Hospital Rebleeding Among Cirrhotic Patients With Acute Esophageal/Gastric Variceal Hemorrhage Using Endoscopy As Gold Standard Maqsood Ahmad, Naeem Afzal, Saeed Akhter, Muhammad Irfan, Zahid Yaseen Hashmie

Abstract The rate of rebleeding of esophageal varices remains observation for re-bleed in 10 days. Results: The high after cessation of acute esophageal variceal mean age of patients was 50.99±12.47 years. Out of hemorrhage. Rockall scoring system primarily 87 positive cases on Rockall score, 82 cases were developed for patients with non variceal source of true positive. Out of 89 cases, that were negative, 81 bleeding. Objective: To determine the validity of were true negative. Sensitivity, Specificity and Rockall scoring system in determining in-hospital diagnostic accuracy of Rockall score was found to re-bleeding among cirrhotic patients with acute be 91.1%, 94.1% and 92.6%, respectively. Positive esophageal/gastric variceal haemorrhage using predictive value (PPV) of Rockall score was 94.2% endoscopy as gold standard. The study was and negative predictive values (NPV) was 90.0%. conducted in Allied hospital and Liver Centre Conclusion: The risk scoring system developed by Allied/DHQ Hospitals, Faisalabad. Study Period: Rockall and coworkers is a clinically useful scoring Study was carried out from 28-06-2010 to 27-12- system among cirrhotic patients with acute 2010. Patients and Methods: Total 176 cases were esophageal/gastric variceal haemorrhage. Key include in this study. The subjects were scored using Words: Esophageal varices, Cirrhosis, Rockall Rockall scoring system and placed under scoring system, Bleeding.

INTRODUCTION Cirrhosis of liver mostly caused by chronic viral are used to stop variceal bleed and to prevent hepatitis is one of the common clinical conditions recurrence 5.Although overall survival has improved in leading to morbidity and mortality in our population. recent years, mortality is still closely related to control Approximately 50% to 60% of patients with cirrhosis of hemorrhage or early re-bleeding, which occurs in as have esophageal varices, and arround 30% of patients many as 40% of patients within first 5 days after initial with varices develop serious bleeding from the varices bleeding episode 6-. Thus the main goal of management within 2 years of diagnosis 1. Despite substantial is to identify patients at high risk for an adverse improvements in the early diagnosis and treatment of outcome on the basis of clinical, laboratory, and variceal hemorrhage, variceal bleeding has a very high endoscopic variables. Scoring system based on these early mortality rate of up to 30% and up to 47%-74% of variables is needed to identify patients who are at risk patients will have recurrent bleeding 2,3 . The reported of re-bleed 7,8 . Rockall scoring system primarily mortality from first episode of variceal bleeding in developed for patients with non variceal source of western studies ranges from 17% to 57%, as compared bleeding and later on applied to variceal bleeding, is to 5-10% mortality reported in our population 4 found to be effective for predicting re-bleeding and Pharmacological therapy, endoscopic intervention and death. The scoring system is based on three clinical other modalities like balloon temponade, and variables (age, shock and co-morbidity) and two radiological and surgical treatment for difficult cases endoscopic variables (diagnosis and major stigmata of

A.P.M.C Vol: 5 No.2 July-December 2011 85

recent hemorrhage, each characterized and scored with DATA COLLECTION 0-3 points, to give maximum score of 11 points as Approval for conduct of study was taken from hospital shown in table 7 A total score of < 2 predicts a low risk ethical committee. The indoor patients with variceal of re-bleeding and mortality. High score (> 8) predicts bleed those fulfilling the inclusion criteria were 8 high risk of re-bleeding and mortality. Endoscopy is recruited for the study. Patients were treated with the gold standard for evaluation of esophageal/gastric pharmacologic therapy i.e. octeriotide, terlipressin and varices 11 . The aim of this study was to evaluate the blood transfusions if needed. Patients were treated with usefulness of Rockall scoring system for stratification of patients with esophageal/gastric varices into high and band ligation and/or injection sclerotherapy as needed. low risk categories for re-bleed during their hospital Informed consent was taken from the patients. In case, stay so that it could also be used to select low risk patient is not capable of giving an informed consent, patients for early discharge and outpatient management. consent was taken from attending guardian. The subjects were scored using Rockall scoring system by MATERIAL AND METHODS the investigator and placed under observation for re- bleed for 10 days. In case of re-bleed the subjects were Study Design treated with pharmacologic therapy i.e. octeriotide, Cross Sectional Validation Study terlipressin, blood transfusions and band ligation Setting and/or injection sclerotherapy as needed. Endoscopy Medical Units and Liver Centre Allied/DHQ Hospitals, was done at the earliest after resuscitation and Faisalabad Duration of Study haemodynamic stability achieved (within 24 hours) by Study was carried out over a period of six months from a consultant endoscopist of our concerned unit. All 28-06-2010 to 27-12-2010. information was recorded before the discharge of Sample Size patients using a structured questionnaire (proforma Sample size was 176 cases attached). Sensitivity and specificity of Rockall score for the detection of rebleed i.e. 91% 9 and 41.1 [9]. DATA ANALYSIS Prevalence of Rebleed: 47% [2, 3] Data was entered and analyzed using SPSS 10.0. Desired Precision: 6% Descriptive statistic was calculated for all variables. Confidence Level: 95% Quantitative variables of the study like age was Sampling Technique expressed as Mean±SD. The sex, findings of Non-probability/ consecutive Sampling ultrasonography and endoscopy findings, rebleed sample selection Sample Selection Rockall score were qualitative variables and presented as frequency and percentage. Validity was assessed by INCLUSION CRITERIA calculating sensitivity, specificity, positive and 1. Age: Patients above the age of 18 years negative predictive values using 2 X 2 tables. 2. Gender: Either ( Male and female) Positive Negative Positive 3. Cirrhotic patients presenting with upper GI bleeding Positive True Positive (a) False Positive (b) as per operational definition confirmed by predictive value Negative Positive False Negative (c) True Negative (d) endoscopy to have esophageal/gastric variceal predictive value source of bleeding. ↓ ↓

Sensitivity Specificity EXCLUSION CRITERIA Sensitivity = a / a + c × 100 1. Cirrhotic patients with non-variceal upper GI bleed Specificity = d / b + d× 100 on endoscopy. Positive Predictive Value = a / a + b × 100 2. Patients with bleeding disorders that are confirmed Negative Predictive Value = d / c + d× 100 by history, clinical examination and low platelet counts, deranged PT and APTT. A.P.M.C Vol: 5 No.2 July-December 2011 86

RESULTS Table-4 A sample of 176 patients was collected from Medical Sensitivity, Specificity and diagnostic accuracy of Units and Liver Centre Allied/DHQ Hospitals, Rockall score Faisalabad, during the study period of six months. The True Positive mean age of patients was 50.99±12.47 years. 43 Sensitivity rate = ______x 100 = patients (24.4%) were in age range of 20-40 years, 105 True Positive + False Negative patients (59.7%) were between age range of 41-60 82 years, 26 patients (14.8%) were 61-80 years while 2 ______x 100 = 91.1% patients (1.1%) were more than 80 years old (Table-1). 82 + 8 Distribution of sex shows, 100 patients (56.9%) were True Negative male while remaining 76 patients (43.1%) were female Specificity rate= ______x 100 (Table-2). Out of 87 positive cases on Rockall score, True Negative + False Positive 82 cases were true positive. Out of 89 cases, that were negative, 81 were true negative (Table-3). Sensitivity, 81 Specificity and diagnostic accuracy of Rockall score ______x 100 = 94.1% 81 + 5 was found to be 91.1%, 94.1% and 92.6%, respectively (Table-4). Positive predictive value (PPV) of Rockall True Positive + True Negative score was 94.2% and negative predictive values (NPV) Diagnostic Accuracy=______x 100 was 90.0% (Table-5). True Positive +True Negative + False Positive + False Negative 82 + 81 Table-1 ______x 100 = 92.6% Distribution of cases by age 82+81+5+8 Age (Year) Number Percentage 20-40 43 24.4 Table-6 41-60 105 59.7 61-80 26 14.8 Positive predictive value and negative > 80 02 1.1 predictive value of Rockall score Total 176 100.00 True Positive Mean±SD 50.99±12.47 Predictive value of= ______x 100 = Positive test True Positive + False Positive 82 Table-2 ______x 100 = 94.2% Distribution of cases by sex 82 + 5 Sex Number Percentage True Negative Male 100 56.9 Predictive value of = ______x 100 Female 76 43.1 Negative test True Negative + False Negative Total 176 100.00 81 ______x 100 = 90.0% Table-3 81 Comparison of Rockall score vs Endoscopy n = 100 DISCUSSION Endoscopy The upper gastrointestinal haemorrhage remains a Rockall score (Gold Standard) Total significant cause of hospital admission, with mortality Positive Negative rates up to 14%. In order to standardise and improve Positive 82 (TP) 5 (FP) 87 care, various scoring systems (e.g. Rockall, Blatchford Negative 08 (FN) 81 (TN) 89 and Baylor) have been developed to identify those Total 90 86 176 individuals at high risk of requiring treatment Key: (transfusion, endoscopic or surgical intervention) or of TP = True positive re-bleeding or death. There is also increasing interest FP = False positive in the utilisation of scoring systems to identify FN = False negative individuals at low risk of complications, as these may TN = True negative A.P.M.C Vol: 5 No.2 July-December 2011 87

be discharged early, possibly with outpatient variceal bleeding. J Coll Physicians Surg Pak endoscopy. Most scoring systems are developed to 2007;17:253-6 predict outcomes in non-variceal bleeding 10 . The 7. Bessa X, O'Callaghan E, Balleste B, Nieto M, Rockall coring system was developed from an analysis Seoane A, Panades A et al. Applicability of of data obtained from a large audit of patients who Rockall score in patients undergoing endoscopic presented with upper-GI bleeding because of a wide therapy for upper gastrointestinal. Dig Liver Dis range of causes, and only a minority of patients 2006;38:12- received endoscopic therapy. The current analysis was 8. Barkun A, Sabbah S, Enns R; Armstrong D ; designed to determine in-hospital re-bleeding among Gregor J ; Fedorak RN et al, RUGBE cirrhotic patients with acute esophageal/gastric Investigators. The Canadian Registry on variceal haemorrhage using endoscopy as gold Nonvariceal Upper Gastrointestinal Bleeding and standard. Endoscopy (RUGBE): endoscopic hemostasis and proton pump inhibition are associated with CONCLUSION improved outcomes in a real-life setting. Am J The risk scoring system developed by Rockall and Gastroenterol 2004; 99:1238-46. coworkers is a clinically useful scoring system among 9. Qureshi W, Adler DG, Davila R, Egan J, Hirota cirrhotic patients with acute esophageal/gastric W, Leighton J, et al. ASGE Guideline: the role of variceal haemorrhage. Rockall scoring system has endoscopy in the management of variceal good predictive and discriminative value for in- hemorrhage, updated July 2005. Gastrointest hospital rebleeding. Endosc 2005; 62:651–655. 10. Atkinson RJ. Usefulness of prognostic indices in REFERENCES upper gastrointestinal Bleeding. Clin 1. Sarwar S, Dilshad A, Ahmad A, Alam A, Butt Gastroenterol 2008; 22:233-42. AK, Tariq S, et al. Predictive value of Rockall 11. Church NI, Dallal HJ, Masson J, Mowat AG, score for rebleeding and mortality in patients Johnston DA, Radin E, et al. Validity of the with variceal bleeding. J Coll Physicians Surg Rockall scoring system after endoscopic therapy Pak 2007; 17:253-6. for bleeding peptic ulcer: a prospective cohort 2. Franchis R. Evolving consensus in portal study. Gastrointest Endosc 2006; 63:606-1

hypertension. Report of the Baveno IV consensus workshop on methodology of diagnosis and AUTHORS therapy in portal hypertension. J Hepatol. 2005; • Dr. Maqsood Ahmad 43:167–176. Assistant Professor of Medicine 3. Garcia-Tsao G, Sanyal AJ, Grace ND, Carey W. Punjab Medical College, Faisalabad Prevention and management of gastroesophageal • Dr. Naeem Afzal varices and variceal hemorrhage in cirrhosis. Senior Registrar Medicine Hepatology. 2007; 46:922–38. Services Hospital Lahore 4. Farooqi RJ, Farooqi JI, Rehman M, Ahmad H, • Dr. Saeed Akhter Ahmed F, Gul S. Outcome after injection Consultant Physician sclerotherapy for esophageal variceal bleeding in MMT Hospital Faisalabad patients with liver cirrhosis and COPD. J Posgrad • Dr. Muhammad Irfan Med Inst 2005;19:76-80. Registrar Allied Hospital Faisalabad 5. Tierney LM, McPhee SJ, Papadakis MA. Current • Dr. Imran Abdullah Medical Diagnosis and Treatment. United States SMO PINUM Faisalabad of America: McGraw Hill Companies 2005; • Prof. Dr. Zahid Yaseen Hashmie 555-56. Professor& Head of Medicine Department 6. Sarwar S, Dilshad A, Ahmad A, Alam A, Butt Punjab Medical College Faisalabad AK, Tariq S, et al. Predictive value of Rockall score for rebleeding and mortality in patients with

A.P.M.C Vol: 5 No.2 July-December 2011 88